45
Systematic Serendipity: Finding the Undiscovered Answers to Science Questions presented by Eugene Garfield Chairman Emeritus, ISI President and Founding Editor, The Scientist 3501 Market Street, Philadelphia, PA 19104 Tel. 215-243-2205 / Fax 215-387-1266 E-mail: [email protected] Home Page: http://eugenegarfield.org The Medical Ignorance Collaboratory University of Arizona Health Sciences Center, Tucson When the 1961 Science Citation Index was first published in 1964, Chemical & Engineering News asked Dr. Julian Smith to review the initial five-volume set. i These volumes were the culmination of an experimental project supported by the National Institutes of Health called the Genetics Citation Index. The Genetics Citation Index was distributed to 1,000 geneticists and contained several different indexes including a 15-year cumulative citation index to the American Journal of Human Genetics, a five-year index to a group of 35 genetics journals, and a selective one-year multi-disciplinary citation index based on 600 journals, published in 1961. ii When the GCI project was almost done in 1963, we asked the National Science Foundation to publish and distribute the complete 1961 Science Citation Index. Since neither the traditional indexing and abstracting services nor NSF would consider publishing it, the Institute for Scientific Information made the decision to do so and more importantly, to continue publication on a quarterly and annual basis. The first quarterly issue appeared in May 1964. Since then an untold number of scientists have been making use of the SCI to retrieve information and to conduct statistical, bibliometric, and sociometric studies. After 40 years, I still can’t take for granted that any audience I address is familiar with it. Here in a nutshell is a brief tutorial.

Systematic Serendipity: Finding the Undiscovered Answers ...garfield.library.upenn.edu/papers/az072004.pdf · serendipidity as “the discovery, by chance or sagacity, of valid results

  • Upload
    others

  • View
    0

  • Download
    0

Embed Size (px)

Citation preview

Page 1: Systematic Serendipity: Finding the Undiscovered Answers ...garfield.library.upenn.edu/papers/az072004.pdf · serendipidity as “the discovery, by chance or sagacity, of valid results

Systematic Serendipity: Finding the Undiscovered Answers to Science Questions

presented by

Eugene Garfield Chairman Emeritus, ISI

President and Founding Editor, The Scientist 3501 Market Street, Philadelphia, PA 19104

Tel. 215-243-2205 / Fax 215-387-1266 E-mail: [email protected] Home Page: http://eugenegarfield.org

The Medical Ignorance Collaboratory

University of Arizona Health Sciences Center, Tucson

When the 1961 Science Citation Index was first published in 1964, Chemical & Engineering News asked Dr. Julian Smith to review the initial five-volume set.i These volumes were the culmination of an experimental project supported by the National Institutes of Health called the Genetics Citation Index. The Genetics Citation Index was distributed to 1,000 geneticists and contained several different indexes including a 15-year cumulative citation index to the American Journal of Human Genetics, a five-year index to a group of 35 genetics journals, and a selective one-year multi-disciplinary citation index based on 600 journals, published in 1961.ii When the GCI project was almost done in 1963, we asked the National Science Foundation to publish and distribute the complete 1961 Science Citation Index. Since neither the traditional indexing and abstracting services nor NSF would consider publishing it, the Institute for Scientific Information made the decision to do so and more importantly, to continue publication on a quarterly and annual basis. The first quarterly issue appeared in May 1964. Since then an untold number of scientists have been making use of the SCI to retrieve information and to conduct statistical, bibliometric, and sociometric studies. After 40 years, I still can’t take for granted that any audience I address is familiar with it. Here in a nutshell is a brief tutorial.

Page 2: Systematic Serendipity: Finding the Undiscovered Answers ...garfield.library.upenn.edu/papers/az072004.pdf · serendipidity as “the discovery, by chance or sagacity, of valid results

HOW TO DO A SEARCH -- HAROLD UREY The SCI system consists of several separate but interrelated indexes organized by author, key word, etc. However, the unique feature which inspired its name – the Citation Index -- permits the user to determine where a particular journal article, book, or patent, has been cited. From that simple linking relationship – cited and citing works -- significant capabilities have evolved. A simple mathematical representation of the scientific literature is to say it is a topological map of about 100,000,000 nodes linked together by about two billion vertices. Slide 2 illustrates by example how a paper by Harold Urey in Science about “Lifelike Forms in Meteorites” has been cited by G. Mueller et al. in Nature.

Page 3: Systematic Serendipity: Finding the Undiscovered Answers ...garfield.library.upenn.edu/papers/az072004.pdf · serendipidity as “the discovery, by chance or sagacity, of valid results

Figure 3 :TOPOLOGICAL MAP OF CITING AND CITED PAPERS

Slide 3 illustrates the topological mapping of 26 papers in genetics that have been linked by various citations.

Page 4: Systematic Serendipity: Finding the Undiscovered Answers ...garfield.library.upenn.edu/papers/az072004.pdf · serendipidity as “the discovery, by chance or sagacity, of valid results

Slide 4: MAPPING THE C. ELEGANS INTERACTOME

This topological map was recently distributed with the June 21st edition of The Scientist.

URL of article: http://www.the-scientist.com/yr2004/jun/feature_040621.html

Page 5: Systematic Serendipity: Finding the Undiscovered Answers ...garfield.library.upenn.edu/papers/az072004.pdf · serendipidity as “the discovery, by chance or sagacity, of valid results

Slide 5: “SYSTEMATIC SERENDIPITY,” BY JULIAN F. SMITH. Chemical and Engineering News 42(35): 55-56 (1964)

Back in 1964, an experienced literature searcher, Julian Smith very perceptively understood how a Citation Index could aid retrieval and dissemination of information.iii In particular, he appreciated that this feature would lead to unexpected connections between citing and cited documents. In a traditional key word search one does not know exactly what will be retrieved but the result is not entirely surprising, that is unexpected, since the documents retrieved contain that key word. On the other hand, when you examine a list of papers that have cited a target article or book, the citing papers may or may not contain the same key words as the target. And due to the multi-disciplinary nature of the SCI’s coverage, one often encounters completely surprising unexpected papers. That explains the origin of the oxymoron “Systematic Serendipity” which Smith used to describe the citation-based search.

Page 6: Systematic Serendipity: Finding the Undiscovered Answers ...garfield.library.upenn.edu/papers/az072004.pdf · serendipidity as “the discovery, by chance or sagacity, of valid results

SLIDE 6: MERTON’S QUOTE AND OED DEFINITION OF SERENDIPITY.

It is of interest to mention Professor Robert K. Merton, the eminent sociologist of science,iv who served on the Advisory Board of the SCI during a 40-year period until his recent death. I recommend his recent and posthumously published book, Travels and Adventures of Serendipity.”v Dr. Merton defined serendipidity as “the discovery, by chance or sagacity, of valid results which were not sought for.”vi The Oxford English Dictionary defines it somewhat differently as “the faculty of making happy and unexpected discoveries by accident,” and properly refers to its first occurrence, indeed invention, by Horace Walpole in 1754. It has been my personal experience that these unexpected serendipitous discoveries have proven most enlightening for users of the Science Citation Index.

Page 7: Systematic Serendipity: Finding the Undiscovered Answers ...garfield.library.upenn.edu/papers/az072004.pdf · serendipidity as “the discovery, by chance or sagacity, of valid results

In 1966, I defined the phenomenon of information retrieval as a dichotomy which includes both information recovery and information discovery. Traditional searches involve information recovery.vii Citation-based searching leads to information discovery. Traditional information experts normally want precision in searching. They prefer not to deal with the unexpected. They measure precision and recall when studying relevance. But relevance, like beauty, is in the eye of the beholder. In spite of numerous large-scale studies of information systems designed to measure precision recall and relevance, especially the work of Cyril Cleverdon,viii it is remarkable that there has never been a comparable study of citation-based retrieval. Relevance has a different meaning in the context of a citation-based search. Only the user can decide what is relevant There is no a priori way to measure that! What does all of this have to do with medical ignorance? If I understand the purpose of this conference, one hopes to demonstrate new mechanisms and technologies to overcome the inherent shortcomings of tradition-based knowledge seeking and searching which I consider a metaphor about education. Ignorance occurs when one relies upon a priori conceptions of relevance, that is, knowledge. Openness to the new and unexpected is in fact a key factor in overcoming ignorance. Professor Witte enticed me to participate in this conference by invoking Robert K. Merton’s name not only because he was my mentor, but also because he published a highly relevant work in 1987 on specified ignorance.ix SLIDE 7: DEFINITION OF SPECIFIED IGNORANCE AND URL FOR MERTON HISTCITE COLLECTION. Robert K. Merton’s Definition of Specified Ignorance: “The express recognition of what is not yet known but needs to be known in order to lay down the foundation for still more knowledge” [HistCite file on the work of Robert K. Merton http://garfield.library.upenn.edu/histcomp/index-merton.html ]

Page 8: Systematic Serendipity: Finding the Undiscovered Answers ...garfield.library.upenn.edu/papers/az072004.pdf · serendipidity as “the discovery, by chance or sagacity, of valid results

He defined specified ignorance as “the express recognition of what is not yet known but needs to be known in order to lay down the foundation for still more knowledge.” The citational impact of that paper and a dozen of Merton’s classic books and papers can be seen at the website I recently created at http://garfield.library.upenn.edu/histcomp/index-merton.html . These HistCite databases are derivatives of the Science, Social Sciences, and Arts and Humanities Citation Indexes, as stored on the Web of Science. Based on my premise that most of you have never used the SCI or SSCI in its printed, CD-ROM, or on-line web formats, here is the absurdly brief tutorial I promised earlier. In the next few slides, I have illustrated the use of SCI to conduct both a quasi-traditional keyword search of a topic as well as a less traditional citation-based search to find information on the topic of medical ignorance. Let me explain what I mean by quasi-traditional. In the SCI, one can indeed do an ordinary title word search. By now title word searching has become traditional in Medline and elsewhere. But when we first used this natural language approach in Current Contents weekly indexes, and later in the SCI, it was pooh-poohed by traditionalists. Controlled vocabulary was considered obligatory. Searching so-called natural language was allegedly fraught with dangers. In the intervening 40 years that bugaboo is now mainly forgotten.x But in Current Contents, we were dealing with the current literature where new terminology cropped up almost daily. And further, we did not suggest that searching titles was comprehensive but rather that it was sufficient as a starting point of further searching. To search current literature one needs current terminology. Slide 8: To poke fun at traditionalists, I often use a quote from Humpty Dumpty in Lewis Carrol’s Through the Looking Glass:

'When I use a word,' Humpty Dumpty said, in a rather scornful tone,' it means just what I choose it to mean, neither more nor less.'

'The question is,' said Alice, 'whether you can make words mean so many different things.'

Page 9: Systematic Serendipity: Finding the Undiscovered Answers ...garfield.library.upenn.edu/papers/az072004.pdf · serendipidity as “the discovery, by chance or sagacity, of valid results

SLIDE 8: HUMPTY DUMPTY SYNDROME

From my experience at the Welch Medical Library Project, I knew that the expert indexers at Index Medicus were very adept at assigning subject headings. After several years in the development of SCI, while relying on natural language searching, we recognized the benefits of human-based indexing, by augmenting our subject search capability by including author supplied descriptors or subject headings. These are now commonplace, but it took many years for the practice to be adopted by scientific and medical journals. We also recognized from the earliest days of SCI’s development that key cited references were in fact symbols for the subject matter they represented.xi Further, the titles of the documents cited in each paper contain a trove of useful descriptive information that can explicate the subject matter discussed in the paper. By assembling and massaging the 25 or so titles of the references cited in a current paper, we were able to create “key words plus.”xii As an essentially automatic procedure KeyWords Plus is used quite effectively to augment title words, author-supplied keywords, or other journal-supplied descriptors. This type of indexing is called “derivative indexing” and it has been successfully used with Medical Subject Headings.xiii Hence my use of the term quasi-traditional.

Page 10: Systematic Serendipity: Finding the Undiscovered Answers ...garfield.library.upenn.edu/papers/az072004.pdf · serendipidity as “the discovery, by chance or sagacity, of valid results

SLIDE 9: RESULTS OF TITLE SEARCH ON “MEDICAL IGNORANCE.”

To find literature on the topic of “Medical Ignorance,” I used the key word approach. In this next slide, I have conducted a “general search” in the SCI on the Web of Science. In Slide 9, we see that there were 9 papers which included the term “medical ignorance” in their titles. Among these are a few papers by Professor Witte. This is a traditional title-based search.

Page 11: Systematic Serendipity: Finding the Undiscovered Answers ...garfield.library.upenn.edu/papers/az072004.pdf · serendipidity as “the discovery, by chance or sagacity, of valid results

SLIDE 10: RESULTS FOR EXPANDED SEARCH ON “MEDICAL IGNORANCE IN TITLES, OR ABSTRACTS, OR KEY WORDS, OR KEYWORDS PLUS.

In Slide 10 , we see the impact of expanding the search to include papers containing “medical ignorance” not only in their titles, but also in key words, or keywords plus, or abstracts. Hence, a quasi-traditional search. This produced 12 papers. By doing this, we have increased recall but lowered precision. SLIDE 11: PAPERS WITH MEDICAL “AND” IGNORANCE IN THEIR TITLES ONLY.

Page 12: Systematic Serendipity: Finding the Undiscovered Answers ...garfield.library.upenn.edu/papers/az072004.pdf · serendipidity as “the discovery, by chance or sagacity, of valid results

Similarly, we could expand the simple title search to look for papers that contain both terms “medical” and “ignorance.” That combination produces one more title.

Page 13: Systematic Serendipity: Finding the Undiscovered Answers ...garfield.library.upenn.edu/papers/az072004.pdf · serendipidity as “the discovery, by chance or sagacity, of valid results

SLIDE 12: PAPERS WITH MEDICAL “AND” IGNORANCE IN THEIR TITLES, ABSTRACTS Similarly, expanding the search to include all key words and the full text of abstracts, produces 141 Papers, most of which in fact are quite distant from medical ignorance.

Page 14: Systematic Serendipity: Finding the Undiscovered Answers ...garfield.library.upenn.edu/papers/az072004.pdf · serendipidity as “the discovery, by chance or sagacity, of valid results

SLIDE 13: PUBMED SEARCH For those of you who are more familiar with PubMed or Medline, the same outcome is possible. In Slide 13, you see the result of a PubMed search. In the previous slides, I illustrated variations on language-based searching. However, once we have identified papers by Dr. Witte and others who have written on this subject, we can quickly augment those results by using the WOS feature called “cited reference” searching.

Page 15: Systematic Serendipity: Finding the Undiscovered Answers ...garfield.library.upenn.edu/papers/az072004.pdf · serendipidity as “the discovery, by chance or sagacity, of valid results

SLIDE 14: FULL RECORD FOR WITTE, KERWIN, ET AL, “A CURRICULUM ON MEDICAL IGNORANCE,” MEDICAL EDUCATION, 23(1):24-29 (JAN, 1989) Slide 14 shows the full source record for one of the Witte papers that was retrieved in the keyword search. By clicking on “times cited,” we find seven additional papers that have cited it.

SLIDE 15: SEVEN PAPERS CITING WITTE 1989 PAPER.

Slide 15 (below) shows the list of seven papers. Note that we are now in the realm of the unexpected. Nothing in the title of the first paper mentions medical ignorance. It is about medical students’ rating of faculty teaching in a multi-instructor setting.

Page 16: Systematic Serendipity: Finding the Undiscovered Answers ...garfield.library.upenn.edu/papers/az072004.pdf · serendipidity as “the discovery, by chance or sagacity, of valid results
Page 17: Systematic Serendipity: Finding the Undiscovered Answers ...garfield.library.upenn.edu/papers/az072004.pdf · serendipidity as “the discovery, by chance or sagacity, of valid results

Moving down the list, one finds three other papers by Witte, but in the first two the term “ignorance” is used, but it was missed in the title search because the noun “medicine” was used rather that the adjective “medical.” The fifth paper is a paper by E. Hearst on “Psychology and Nothing” in the 1991 American Scientist. SLIDE 16: PAPER BY E. HEARST, “PSYCOLOGY AND NOTHING,” IN AMERICAN SCIENTIST. In contrast to the papers by Witte on Medical Ignorance, I would classify this paper as “unexpected.” However, we will see later that it is not completely unrelated I tracked down this paper by locating the author who is right here at the University of Arizona. The subtitle of the paper gives us a small clue to its relevance: “Recognizing and learning from absence, deletion and nonoccurrence are surprisingly difficult. Animals and people, it seems, accentuate the positive.” Then in the penultimate paragraph, the author states: “Scientific training emphasizing disproof rather than confirmation is not as widespread as it probably ought to be. And successful courses of instruction are being developed that stress absence of knowledge (“ignorance”) or examination of past failures, rather than presentation of presumed positive conclusions or outcomes, as their primary educational focus.” He then cites the University of Arizona program on medical ignorance. [Popperian, “resistance to falsification.”] Well, I hope that this brief tutorial on using WOS has illustrated the key point. While using key words is essentially an a priori view of the literature, the use of cited reference searching is essentially a posteriori. While cited and citing works can be closely connected, they can also seem at first unrelated, even serendipic. And in some cases the only thing that may link two papers is the particular method they cite. While this may be uninteresting to one reader, it is quite beautiful to those interested in the methodology involved. I could also go on to describe bibliographic coupling which ISI calls “related records.” For example, in Slide 16 by clicking on the related records button you find the following list of related, that is, bibliographically coupled papers.

Page 18: Systematic Serendipity: Finding the Undiscovered Answers ...garfield.library.upenn.edu/papers/az072004.pdf · serendipidity as “the discovery, by chance or sagacity, of valid results

Slide 16.

Page 19: Systematic Serendipity: Finding the Undiscovered Answers ...garfield.library.upenn.edu/papers/az072004.pdf · serendipidity as “the discovery, by chance or sagacity, of valid results

SLIDE 17: RELATED RECORDS FOR HEARST These papers are ranked by the number of references that they share in common with the Hearst paper on “Psychology and Nothing.” No less than 3,102 papers are related in this way but in the next slide we see the top ten which give some idea of what that topic may be about.

Page 20: Systematic Serendipity: Finding the Undiscovered Answers ...garfield.library.upenn.edu/papers/az072004.pdf · serendipidity as “the discovery, by chance or sagacity, of valid results

SLIDE 18: RELATED RECORDS TO WITTE

As a further illustration of related record searching, let’s return to Witte’s paper in Slide 14 and we find that there are 1,383 papers with connecting links but the top eight are shown in Slide 18. You can be the judge whether these papers are or are not related to “a curriculum on medical ignorance.” We have all heard the term “information overload.” That term was used long before the advent of the internet, but the internet has aggravated the problem. By removing the physical barriers to traditional manual-based searching of databases, the process of current awareness has been made more efficient but we still bump up against the limitation of time for reading. It is not unusual when doing a literature search today to encounter a volume of relevant literature that is impossible to read in its entirety. It is true whether you are using Medline, Google, SCI, or any other search engine. The problem, therefore, is how we separate the wheat from the chaff.

Page 21: Systematic Serendipity: Finding the Undiscovered Answers ...garfield.library.upenn.edu/papers/az072004.pdf · serendipidity as “the discovery, by chance or sagacity, of valid results

As I have indicated earlier, there are two aspects to information retrieval. One is information recovery – the other information discovery. When we first published the SCI, we quickly realized that machine readable files could be the basis of a system for selective dissemination of information (SDI).xiv Just one year after we inaugurated the primordial SCI, ISI became the first organization to offer a commercial system for the selective dissemination of information (SDI). This took expression in the form of a weekly computer printout that told scholars of current papers that cited their work or the work of colleagues. I have been the beneficiary of this service for almost 40 years. It still amazes me how often I am made aware of papers that I could not have anticipated by the use of key words. Similarly, Dr. Merton also received thousands of notifications over the years. While a highly-cited scholar, the personal alerting system produced relatively small weekly listings. So in spite of how often he was cited, it was quite feasible for him to scan those weekly lists and decide which papers warranted reprint requests. Today one can increasingly find full texts on the internet or identify the email address of authors. This will become routine as time goes by. You can even obtain this kind of alert service using the Google system. Those of you who use the Google News Alert will recognize the similarity between it and the ISI Alert. The WOS system I described above also has an alerting feature. In one form or another, the problem of information overload has been recognized from the earliest days of scientific publishing. After the invention of printing in the 15th Century, scholars were soon complaining that it was impossible to read everything published. And especially from the end of the 19th Century to this day, bibliographers and librarians have proposed alternative solutions. Not the least of these were modern classification systems like the Universal Decimal System (UDC) on which is based on the Dewey Decimal System. However, it was Henry E. Bliss who designed the first classification for science. When I met him in 1954, I was a student at Columbia Library School. He told me that a significant objective of classification is to help the reader see the connections between subject matter. To create his classification system, he read the textbook in every known branch of science over a 40-year period as librarian of CCNY – the City College of the City University of New York, an institution that has produced many scholars of Nobel Class. Another method showing the connections between topics is implicit in the growing field of visualization. Back in the sixties my colleagues and I described a technique for visualizing topics chronologically. We called them Historiographs. In 1964, we issued a report "The Use of Citation Data in Writing the History of Science."xv

Page 22: Systematic Serendipity: Finding the Undiscovered Answers ...garfield.library.upenn.edu/papers/az072004.pdf · serendipidity as “the discovery, by chance or sagacity, of valid results

SLIDE 19: HISTORIOGRAPH: FROM MENDEL TO NIRENBERG Out of this came an historiograph of the literature from Gregor Mendel (1865) to Avery and McLeod (1944) to Watson-Crick (1953) and Nirenberg (1961). We used IBM punched card methods to sort the data and drew our maps manually.

Page 23: Systematic Serendipity: Finding the Undiscovered Answers ...garfield.library.upenn.edu/papers/az072004.pdf · serendipidity as “the discovery, by chance or sagacity, of valid results

SLIDE 20: PUNCHED CARD In case you have never seen a punched card, here are some examples. Since then, we have automated the procedure for creating historiographs.

Page 24: Systematic Serendipity: Finding the Undiscovered Answers ...garfield.library.upenn.edu/papers/az072004.pdf · serendipidity as “the discovery, by chance or sagacity, of valid results

Slide 21:

Page 25: Systematic Serendipity: Finding the Undiscovered Answers ...garfield.library.upenn.edu/papers/az072004.pdf · serendipidity as “the discovery, by chance or sagacity, of valid results

SLIDE 21: HISTCITE GUIDE The HistCite program takes the output of a search on the ISI Web of Science (the internet version of the SCI) and generates a chronological historiograph of the key literature. Whatever topic is chosen one can simply begin a WOS search with a key term, save a text file of retrieved papers that use that term, and add additional papers that cite them to obtain a group of past and present papers, that are linked to varying degrees. I would like now to illustrate the HistCite system. It has a dual function -- one to help separate the wheat from the chaff in a large collection of papers however they are retrieved by key word or by citation based search, and secondly, to help scholars prepare micro-histories of those topics. Those of you who are familiar with Google and other search engines may recognize some similarities in what I have described earlier. The connection between citation indexing and googling is well recognized by experts.xvi However, it is also well known that most search engines produce too much information. The challenge is to identify the needle in the haystack.

SLIDE 22: PAPERS FOUND IN WOS - TOPIC SEARCH: IGNORAN* AND MEDIC* - CHRONOLOGICAL FILE http://garfield.library.upenn.edu/histcomp/medic_ignoran/

Page 26: Systematic Serendipity: Finding the Undiscovered Answers ...garfield.library.upenn.edu/papers/az072004.pdf · serendipidity as “the discovery, by chance or sagacity, of valid results

Slide 22: LCS http://garfield.library.upenn.edu/histcomp/medic_ignoran/index-lcs.html

Missing Links? Citation Matrix Glossary HistCite Guide AboutPapers found in WoS

Topic Search: ignoran* and medic*

Nodes: 228 Authors: 483, Journals: 179, Outer References: 6842 Collection span: 1961 - 2004 View: Overview. Sorted by year, source, volume, issue, page.

# LCR NCR Nodes / Authors / Date LCS GCS

1 0 0 1 1961 JAMA-JOURNAL OF THE AMERICAN MEDICAL ASSOCIATION 176(9):804-& [Anon] Advanced Medical Ignorance

0 0

2 0 10 2 1962 AMERICAN SOCIOLOGICAL REVIEW 27(4):508-514 LEWIS LS; LOPREATO J Arationality, Ignorance, and Perceived Danger in Medical Practices

0 14

3 0 0 3 1974 SOUTH AFRICAN MEDICAL JOURNAL 48(17):717-718 SMIT PJ Sports Medicine, Ignorance and Sport

0 2

4 0 0 4 1979 AUSTRALIAN AND NEW ZEALAND JOURNAL OF MEDICINE 9(5):629-629 RYAN R Anxious Moments in Nuclear-Medicine - Is Ignorance Bliss

0 0

5 0 6 5 1980 ET CETERA 37(1):5-12 BATES EM The Implications for Semantics of Developments in Medical Technology - The Fatal Consequences of Semantic Ignorance

0 0

6 0 1 6 1984 LANCET 1(8382):884-884 NORTON A The Encyclopedia of Medical Ignorance - Exploring the Frontiers of Medical Knowledge - Duncan,R, Westonsmith,M

0 0

7 0 1 7 1984 NATURE 310(5974):257-257 WEATHERALL DJ The Encyclopedia of Medical Ignorance - Exploring the Frontiers of Medical Knowledge - Duncan,R, Westonsmith,M

0 0

8 0 1 8 1985 NEW YORK TIMES BOOK REVIEW (MAY):33-33 LEFANU J The Encyclopedia of Medical Ignorance - Exploring the Frontiers of Medical Knowledge - Duncan,R, Westonsmith,M

0 0

9 0 16 9 1986 LANGUAGE AND STYLE 19(1):11-20 KELLY LG Medicine, Learned Ignorance, and Style in 17th-Century Translation

0 2

10 0 1 10 1987 LIBRARY QUARTERLY 57(2):235-237 MATHESON NW Consensus and Penalties for Ignorance in the Medical Sciences - Implications for Information-Transfer - Brittain,JM

0

Page 27: Systematic Serendipity: Finding the Undiscovered Answers ...garfield.library.upenn.edu/papers/az072004.pdf · serendipidity as “the discovery, by chance or sagacity, of valid results

By clicking on LCS, we obtain the file sorted by Local Citation Score. The Witte paper in Journal of Medical Education comes to the top In the above slide, I have illustrated the format of a HistCite collection by searching WOS to find papers on “medical ignorance.” SLIDE 23: LCS

SLIDE 23: LCS

Page 28: Systematic Serendipity: Finding the Undiscovered Answers ...garfield.library.upenn.edu/papers/az072004.pdf · serendipidity as “the discovery, by chance or sagacity, of valid results

SLIDE 24: GCS http://garfield.library.upenn.edu/histcomp/medic_ignoran/index-tc.html

The file is now sorted by Global Citation Score. The paper by Brady comes to the top.

Page 29: Systematic Serendipity: Finding the Undiscovered Answers ...garfield.library.upenn.edu/papers/az072004.pdf · serendipidity as “the discovery, by chance or sagacity, of valid results

SLIDE 25: AUTHORS http://garfield.library.upenn.edu/histcomp/medic_ignoran/list/au-pubs.html By clicking on Authors, we find the most published authors in the file.

SLIDE 25: AUTHORS http://garfield.library.upenn.edu/histcomp/medic_ignoran/list/au-pubs.html

Page 30: Systematic Serendipity: Finding the Undiscovered Answers ...garfield.library.upenn.edu/papers/az072004.pdf · serendipidity as “the discovery, by chance or sagacity, of valid results

SLIDE 26: JOURNALS

http://garfield.library.upenn.edu/histcomp/medic_ignoran/list/so-pubs.html A sort by source journal shows that 8 articles on this topic have appeared in a French journal, followed by 6 in Social Science and Medicine.” The rest are widely dispersed in 179 different journals.

Page 31: Systematic Serendipity: Finding the Undiscovered Answers ...garfield.library.upenn.edu/papers/az072004.pdf · serendipidity as “the discovery, by chance or sagacity, of valid results

SLIDE 27: ARTICLES PUBLISHED IN SCIENCES ET TECHNIQUES DE L’ANIMAL DE LABORATORIE I was curious as to why this French veterinary journal would be so prominent. It turns out that the entire group of 8 articles is an eight-part series about vascular microsurgery in the rat. The same abstract is repeated in all of the listings. The relevant line is underlined.

Page 32: Systematic Serendipity: Finding the Undiscovered Answers ...garfield.library.upenn.edu/papers/az072004.pdf · serendipidity as “the discovery, by chance or sagacity, of valid results

SLIDE 28: ARTICLES PUBLISHED IN SOCIAL SCIENCE & MEDICINE On the other hand, in Slide 28 we see the six papers for Social Science & Medicine. Only one of these papers contains the term ignorance in the title while others contain the term only in the abstract. This example illustrates the semantic problems inherent in natural language searches. Nevertheless, the paper, by Maynard has a beautiful quote, “The price of knowledge is high but the cost of ignorance is greater.”

Page 33: Systematic Serendipity: Finding the Undiscovered Answers ...garfield.library.upenn.edu/papers/az072004.pdf · serendipidity as “the discovery, by chance or sagacity, of valid results

SLIDE 29: OUTER REFERENCES http://garfield.library.upenn.edu/histcomp/medic_ignoran/index-tc.html# The “outer references” is a ranked list of everything else that is cited in the main collection. The papers by Eisenberg, Arrow, and Astin appear at the top. Most of the outer references do not use the term ignorance in their titles. That is why they are not in the main collection. An exception is the reference to Duncan’s Encyclopedia of Medical Ignorance. As a book, it would not be a source entry in WOS. To include a more complete sample of relevant papers, we added the top 10 of the outer references to the main collection.

Page 34: Systematic Serendipity: Finding the Undiscovered Answers ...garfield.library.upenn.edu/papers/az072004.pdf · serendipidity as “the discovery, by chance or sagacity, of valid results

SLIDE 30: OUTER REFERENCES ADDED TO MAIN COLLECTION In Slide 30, you see that these additional papers are included, but we have not included the many other global papers that cite them. We have also added some additional papers that have cited the work of Witte on medical ignorance.

Page 35: Systematic Serendipity: Finding the Undiscovered Answers ...garfield.library.upenn.edu/papers/az072004.pdf · serendipidity as “the discovery, by chance or sagacity, of valid results

SLIDE 31: FILE SORTED BY LCS SHOWING WITTE 1988 TO 1990 In Slide 31, the file is sorted by local citation score and shows the 1988 to 1990 Witte papers are at the top, followed by Eisenberg in JAMA, Arrow in American Economic Review, then Sutherland and then Astin.

Page 36: Systematic Serendipity: Finding the Undiscovered Answers ...garfield.library.upenn.edu/papers/az072004.pdf · serendipidity as “the discovery, by chance or sagacity, of valid results

SLIDE 32: GLOBAL CITATION SORT In this slide we see the same file sorted by Global Citation Score. Note that Eisenberg, Arrow, and Astin are Citation Classics, by which I mean they have been cited hundreds of times. I have recently posted a file of 4,000 author commentaries on Citation Classics at http://garfield.library.upenn.edu/classics.html. These originally appeared in Current Contents over a 25-year period. Time does not permit me to take this particular analysis any further.

Page 37: Systematic Serendipity: Finding the Undiscovered Answers ...garfield.library.upenn.edu/papers/az072004.pdf · serendipidity as “the discovery, by chance or sagacity, of valid results

SLIDE 33: HISTORIOGRAPH OF MEDICAL IGNORANCE The detailed analyses we have done so far have demonstrated that the subject of medical ignorance is not a precisely defined subject. There appears to be few papers that clearly epitomize this subject. I also tested the subject “absence of knowledge,” and it produced almost nothing relevant.

Page 38: Systematic Serendipity: Finding the Undiscovered Answers ...garfield.library.upenn.edu/papers/az072004.pdf · serendipidity as “the discovery, by chance or sagacity, of valid results

Let me close the discussion of HistCite by referring to a topic that is rather precisely laid out – the 1953 Watson-Crick paper on the double helix structure of DNA and the controversy concerning its link to the work of Avery-McCleod in 1944. This is discussed in detail in a recent paper I published in an American Library Association journal called Information Technology and Libraries.xvii Now I would briefly like to show you the use of HistCite in a truly historical mapping exercise. We are celebrating the 50th anniversary of the Watson-Crick discovery of the Double Helix structure of DNA. That 1953 paper was used to conduct a cited reference of the SCI. Since time is short, I am not going to dwell on the details of that search. In the following slides, I simply want to show you the result of mapping the five years from 1953 to 1958. SLIDE 34: 1953-1958 ARTICLES THAT CITED WATSON-CRICK 1953 PAPER ON “MOLECULAR STRUCTURE OF DNA” Figure 34 shows the chronological HistCite table for the papers that cite the Watson-Crick, 1953 paper in Nature. We have also added a few of the key outer references for Avery and Hershey. Contrary to claims made in the 50th anniversary issue of Nature, this paper was cited over 200 times during the five years following publication. The papers by Avery and Hershey were outer references in the first iteration of this search. In other word, they were highly cited by the papers that cite the Watson-Crick paper. The Avery paper was cited 150 times during the five-year period. Since then it has been cited in about 1400 papers.

Page 39: Systematic Serendipity: Finding the Undiscovered Answers ...garfield.library.upenn.edu/papers/az072004.pdf · serendipidity as “the discovery, by chance or sagacity, of valid results
Page 40: Systematic Serendipity: Finding the Undiscovered Answers ...garfield.library.upenn.edu/papers/az072004.pdf · serendipidity as “the discovery, by chance or sagacity, of valid results

SLIDE 35: TWO GENERATIONS OF CITATIONS TO WATSON-CRICK 1953 PAPER This HistCite file contains 975 papers that either cited Watson-Crick 53 or cited one of the 210 citing papers. Thus, we have two generations of citations.

Page 41: Systematic Serendipity: Finding the Undiscovered Answers ...garfield.library.upenn.edu/papers/az072004.pdf · serendipidity as “the discovery, by chance or sagacity, of valid results

SLIDE 36: WATSON-CRICK YEAR-BY-YEAR HISTORIOGRAPH 1936-1965

In Figure 36, we have the year-by-year map of the 22 most-cited papers in the chained indexed file. Notice that in 1953 there were nine highly-cited papers. And in 1954, there are five. Using the typical reference citation, that is, only author, volume, page, and year, it is not possible to differentiate the month-by-month progression, after the Watson and Crick paper appeared in April, 1953. However, the HistCite system can take into account the cover dates of the journals involved if they are included. The Watson-Crick paper is #27. Avery 1944 is #2.

SLIDE 37: WATSON-CRICK MONTH BY MONTH 1953 In Figure 37, we see how the historiography changes not just year-to-year but month-by-month. Unfortunately, WoS does not contain cover dates until 1985 or there about. So we had to manually insert in the export files the cover dates for the few dozen papers involved in this example

Page 42: Systematic Serendipity: Finding the Undiscovered Answers ...garfield.library.upenn.edu/papers/az072004.pdf · serendipidity as “the discovery, by chance or sagacity, of valid results

.

Page 43: Systematic Serendipity: Finding the Undiscovered Answers ...garfield.library.upenn.edu/papers/az072004.pdf · serendipidity as “the discovery, by chance or sagacity, of valid results
Page 44: Systematic Serendipity: Finding the Undiscovered Answers ...garfield.library.upenn.edu/papers/az072004.pdf · serendipidity as “the discovery, by chance or sagacity, of valid results

References: 1. Smith JF. “Systematic Serendipity,” Chemical & Engineering News 42(35):55-56 (1964). 2. Garfield, E. and Sher, I.H. "The Genetics Citation Index - Introduction" (In: Eds, Eugene Garfield and Irving H. Sher, Genetics Citation Index, pgs. x-iii, Philadelphia Institute for Scientific Information (1963) 864 pages. Reprinted in Garfield E., Essays of an Information Scientist, Volume 7, pgs. 517-520. Philadelphia: ISI Press (1985). 3. Smith JF. “Systematic Serendipity,” Chemical & Engineering News 42(35):55-56 (1964). See: http://www.garfield.library.upenn.edu/merton/list.html 4. Merton, RK and Barber, EG. Viaggi e avventure della Serendipity. Saggio di semantica sociologica e di sociologia della scienza, Bologna: il Mulino 416 pp. (2002). Published in English by Princeton University Press: The Travels and Adventures of Serendipity: A Study in Historical Semantics and the Sociology of Science. 328 pp. (2004) 5. Merton RK. “Sociological Theory,” American Journal of Sociology 50(6):469 (1945) 6. Garfield, E. “The who and why of ISI -- ISI eases scientists' information problems; provides

convenient orderly access to literature,” Karger Gazette No. 13, pgs 2 (March 5,1966). http://www.garfield.library.upenn.edu/essays/V1p033y1962-73.pdf 7. Cleverdon CW. “Inverse Relationship of Recall and Precision,” Journal of Documentation 28(3):195 (1972). 8. Merton RK., “Three Fragments from a Sociologist’s Notebooks: Establishing the Phenomenon, Specified Ignorance, and Strategic Research Materials,” Annual Review of Sociology 13:1-28 (1987) 9. Jenuwine, ES and Floyd, JA. “Comparison of Medical Subject Headings and Text-Word Searches in MEDLINE to Retrieve Studies on Sleep in Healthy Individuals,” Journal of the Medical Library Association 82(3): 349-353 (July 2004). 10. Small, HG. “Cited Documents as Concept Symbols,” Social Studies of Science 8(3):327-40 (1978) 11. Garfield, E and Sher, IH. " KeyWords Plus -- Algorithmic Derivative Indexing," Journal of the American Society for Information Science, 44 (5):298-299 (1993). http://www.garfield.library.upenn.edu/papers/jasis44(5)p298y1993.html 12. Gray, WA, & Harley, AJ. ”Computer Assisted Indexing,” .Information Storage & Retrieval, 7, 367-174. (1971)

Page 45: Systematic Serendipity: Finding the Undiscovered Answers ...garfield.library.upenn.edu/papers/az072004.pdf · serendipidity as “the discovery, by chance or sagacity, of valid results

13. Garfield, E. and Sher. I. "ASCA (Automatic Subject Citation Alert) -- A New Personalized Current Awareness Service for Scientists ," American Behavioral Scientist, 10 (5) pgs. 29-32 (1967). Reprinted in Essays of an Information Scientist, Volume 6, pgs 514-17. Philadelphia: ISI Press (1984). http://www.garfield.library.upenn.edu/essays/v6p514y1983.pdf 14. Garfield, E., Sher I.H. , & Torpie, R, H. "The Use of Citation Data in Writing the History of Science." Philadelphia: The Institute for Scientific Information, Report of research for Air Force Office of Scientific Research under contract F49(638)-1256 (1964). Available: http://www.garfield.library.upenn.edu/papers/useofcitdatawritinghistofsci.pdf’ 15. Kleinberg JM. “Authoritative sources in a hyperlinked environment,” Journal of the ACM, 46 (5):604-632 (September, 1999) 16. Garfield E, Pudovkin AI, Istomin VI, "Mapping the Output of Topical Searches in the Web of Knowledge and the case of Watson-Crick," Information Technology and Libraries; 22 (4): 183-187 December 2003. Available: http://garfield.library.upenn.edu/papers/casewatsoncrick2003.pdf